Who needs expert systems
What is an expert system (ES)? Everyone has heard this term, it is often perceived as outdated, unfashionable, far from the mainstream in science and technology. "This is a very special area of ​​programming that does not have widespread use." ES are held at the institute and then safely forgotten.
They appeared in the late 60s as the most promising, practical direction in the development of the science of artificial intelligence. Despite many interesting developments, ES are not only not widely spread, but there is not even a single commercially successful development.
If we analyze the concepts of this area of ​​computer science - ES, it becomes clear that, although the direction has not received a noticeable development, the problems here lie at the very basis of the development of computers.
')
When the first electronic computers (computers) appeared, they were called the electronic brain and were wary of the arrival of the machines replacing the human being. There were a lot of machines, their capacity increased millions of times, but they could not replace a person. The appearance of the "electronic brain" in the coming years is not predicted, despite the fact that more than 60 years have passed since the appearance of the first computers.
One of the main reasons for the poor development of ES is the algorithmic paradigm of thinking of application developers. This paradigm has penetrated so deeply into everything that concerns computers that we do not notice it. It seems that everything is great - a person gives commands (presses the necessary buttons on the screen with a mouse) - the machine executes, what more could you ask for? If you dig deeper - what commands does a person give? Instead of giving the car a task - saying WHAT it is necessary to do, what the user wants, the person gives instructions - HOW to do, what sequence of operations to execute to the machine in order to get the desired result.
The first problem in communicating the user with the machine is to learn, learn how to use the necessary application programs — that is, learn when and what buttons in the program to press to get the desired result. Like any car, you need to master it, understand how to handle it, remember control methods and master them, as well as with a car or a coffee maker.
The inconsistency of this situation, no one notices. A computer is not just another machine to make life easier for a person, a computer is a machine for enhancing intelligence, for performing mental operations. (As it has already been forgotten, arithmetic operations and procedures are mental operations that exist only in the mind of a person). Since modern computers perform only algorithmic operations, the rest of the mental activity remains with the person.
As a result, the more programs a person must apply, the more he must memorize and learn. The computer does not "unload", but "loads."
Instead of a skilled, intelligent, knowledgeable “helper” who can be assigned work without worrying about the result, we have a strong and stupid “slave” who will only do what is ordered and not one iota more, who has neither diligence nor knowledge nor creative striving for results.
In order to achieve a result from it, the person - the user must work hard - learn the program, constantly manage it, prepare the initial data for the program and still process the results of the work.
Take, for example, electronic accounting. Accounting programs can be compared with the accountant - a specialist of lower qualification in the hierarchy of accounting specialties. The accounting program can not replace the accountant, it replaces only the accountant - fills in the graphs in the journals and calculates the total amounts. Electronic accounting has many useful properties, but without an accountant it is useless.
What could be an intellectual program - accounting ES? Such a program completely replaces a specialist accountant - automatically and accurately classifies and encodes monetary transactions, knows, automatically collects and applies all necessary regulatory and legal acts and can advise the director on current operations and general accounting issues.
If you take our text editors - they facilitate the work of typists, eliminating the need to reprint the page when errors are detected. As a result, we stopped writing with a pen and became our own typists. Is there any gain in productivity?
The intelligent text editor should create the texts of the documents according to the user's instructions. For example: “You must agree to this letter”, “You must prepare an objection to this claim”, “This letter must be rewritten in a softer form”.
And if you take spreadsheets, the powerful spreadsheet paradigm itself could be much more massive and effective if it were not for the complexity of its development. Most users do not use half of the functions of spreadsheets and text editors due to the difficulty of fully mastering these programs.
Any modern application program can be supplemented or replaced with ES for a drastic increase in the efficiency of computer use. If you move from the processing of symbols and numbers to the processing of semantic information, you can get a real helper - a semantic computer.
It is enough to replace the traditional help with a knowledge base on the use of the program to significantly increase efficiency and free the user from the stress that arises when working with a new program.
One of the important differences between the ES is the user's language orientation, since in the ES the user formulates a description of his task - WHAT he wants, and not instructions to the machine - HOW to complete his task.
There is one gap in modern programming because of which there is no installation in the consciousness of programmers on semantic information processing - there is no formal basis - programming languages ​​for representing and processing semantic information, although all the necessary language tools are described in the relevant publications. The conceptual programming system is being developed in the Semantiks Rishech NPF.
The inertia of the algorithmic approach hinders the creation of mass-scale applied ES. Much can be done on the existing technological base. It is only necessary to remember that the machine is a means of enhancing the human intellect and can perform not only arithmetic operations.
Replacing existing applications with expert systems will dramatically increase the efficiency of computers and improve the modern world. We will give just a few possible examples.
Creating an ES of mass use could greatly increase labor productivity in all areas of human activity.
The EC for a lawyer could assist him in compiling and analyzing documents, advising on current legislation.
ES can fundamentally change the nature of school and university education, as well as significantly improve the usefulness of e-government. The ES tax inspector could automatically check tax returns and other documents.
In the not very distant perspective of the development of the ES, it is possible to imagine that everyone will be able to get unlimited possibilities of interested and active personal communication, which today ensure the commercial success of social networks and Internet search programs. Communication with virtual intellectual personalities will fundamentally change not only the human-computer paradigm of communication, but also the very foundations of the economic activity of society.
They appeared in the late 60s as the most promising, practical direction in the development of the science of artificial intelligence. Despite many interesting developments, ES are not only not widely spread, but there is not even a single commercially successful development.
If we analyze the concepts of this area of ​​computer science - ES, it becomes clear that, although the direction has not received a noticeable development, the problems here lie at the very basis of the development of computers.
')
When the first electronic computers (computers) appeared, they were called the electronic brain and were wary of the arrival of the machines replacing the human being. There were a lot of machines, their capacity increased millions of times, but they could not replace a person. The appearance of the "electronic brain" in the coming years is not predicted, despite the fact that more than 60 years have passed since the appearance of the first computers.
One of the main reasons for the poor development of ES is the algorithmic paradigm of thinking of application developers. This paradigm has penetrated so deeply into everything that concerns computers that we do not notice it. It seems that everything is great - a person gives commands (presses the necessary buttons on the screen with a mouse) - the machine executes, what more could you ask for? If you dig deeper - what commands does a person give? Instead of giving the car a task - saying WHAT it is necessary to do, what the user wants, the person gives instructions - HOW to do, what sequence of operations to execute to the machine in order to get the desired result.
The first problem in communicating the user with the machine is to learn, learn how to use the necessary application programs — that is, learn when and what buttons in the program to press to get the desired result. Like any car, you need to master it, understand how to handle it, remember control methods and master them, as well as with a car or a coffee maker.
The inconsistency of this situation, no one notices. A computer is not just another machine to make life easier for a person, a computer is a machine for enhancing intelligence, for performing mental operations. (As it has already been forgotten, arithmetic operations and procedures are mental operations that exist only in the mind of a person). Since modern computers perform only algorithmic operations, the rest of the mental activity remains with the person.
As a result, the more programs a person must apply, the more he must memorize and learn. The computer does not "unload", but "loads."
Instead of a skilled, intelligent, knowledgeable “helper” who can be assigned work without worrying about the result, we have a strong and stupid “slave” who will only do what is ordered and not one iota more, who has neither diligence nor knowledge nor creative striving for results.
In order to achieve a result from it, the person - the user must work hard - learn the program, constantly manage it, prepare the initial data for the program and still process the results of the work.
Take, for example, electronic accounting. Accounting programs can be compared with the accountant - a specialist of lower qualification in the hierarchy of accounting specialties. The accounting program can not replace the accountant, it replaces only the accountant - fills in the graphs in the journals and calculates the total amounts. Electronic accounting has many useful properties, but without an accountant it is useless.
What could be an intellectual program - accounting ES? Such a program completely replaces a specialist accountant - automatically and accurately classifies and encodes monetary transactions, knows, automatically collects and applies all necessary regulatory and legal acts and can advise the director on current operations and general accounting issues.
If you take our text editors - they facilitate the work of typists, eliminating the need to reprint the page when errors are detected. As a result, we stopped writing with a pen and became our own typists. Is there any gain in productivity?
The intelligent text editor should create the texts of the documents according to the user's instructions. For example: “You must agree to this letter”, “You must prepare an objection to this claim”, “This letter must be rewritten in a softer form”.
And if you take spreadsheets, the powerful spreadsheet paradigm itself could be much more massive and effective if it were not for the complexity of its development. Most users do not use half of the functions of spreadsheets and text editors due to the difficulty of fully mastering these programs.
Any modern application program can be supplemented or replaced with ES for a drastic increase in the efficiency of computer use. If you move from the processing of symbols and numbers to the processing of semantic information, you can get a real helper - a semantic computer.
It is enough to replace the traditional help with a knowledge base on the use of the program to significantly increase efficiency and free the user from the stress that arises when working with a new program.
One of the important differences between the ES is the user's language orientation, since in the ES the user formulates a description of his task - WHAT he wants, and not instructions to the machine - HOW to complete his task.
There is one gap in modern programming because of which there is no installation in the consciousness of programmers on semantic information processing - there is no formal basis - programming languages ​​for representing and processing semantic information, although all the necessary language tools are described in the relevant publications. The conceptual programming system is being developed in the Semantiks Rishech NPF.
The inertia of the algorithmic approach hinders the creation of mass-scale applied ES. Much can be done on the existing technological base. It is only necessary to remember that the machine is a means of enhancing the human intellect and can perform not only arithmetic operations.
Replacing existing applications with expert systems will dramatically increase the efficiency of computers and improve the modern world. We will give just a few possible examples.
Creating an ES of mass use could greatly increase labor productivity in all areas of human activity.
The EC for a lawyer could assist him in compiling and analyzing documents, advising on current legislation.
ES can fundamentally change the nature of school and university education, as well as significantly improve the usefulness of e-government. The ES tax inspector could automatically check tax returns and other documents.
In the not very distant perspective of the development of the ES, it is possible to imagine that everyone will be able to get unlimited possibilities of interested and active personal communication, which today ensure the commercial success of social networks and Internet search programs. Communication with virtual intellectual personalities will fundamentally change not only the human-computer paradigm of communication, but also the very foundations of the economic activity of society.
Source: https://habr.com/ru/post/396155/
All Articles